Generating a key-image from a medical image

Patent metadata

Field	Value
Publication number	US-9697598-B2
Application number	US-201314437839-A
Country	US
Kind code	B2
Filing date	Nov 12, 2013
Priority date	Nov 23, 2012
Publication date	Jul 4, 2017
Grant date	Jul 4, 2017

Abstract

A system ( 100 ) for generating a key-image from a medical image, the key-image comprising an image area selected from the medical image, the system comprising: —an image analyzer ( 140 ) for: i) detecting a lesion ( 300 ) in the medical image; ii) based on anatomical information, identifying an anatomical landmark ( 310 ) which has an anatomical relation with the lesion; and iii) detecting the anatomical landmark in the medical image; and —a key-image generator ( 160 ) for j) selecting an image area ( 400 ) in the medical image which comprises the lesion ( 300 ) and the anatomical landmark ( 310 ), and jj) generating the key-image ( 162 ) so as to comprise said image area.

First claim

The invention claimed is:

1. A system for generating a key-image from a medical image, the key-image comprising an image area selected from the medical image, the system comprising:
an image interface for obtaining the medical image and metadata indicative of a content of the medical image;
a database interface for accessing anatomical information on a medical database, the anatomical information being indicative of anatomical landmarks associated with the content of the medical image;
an image analyzer for:
i) detecting a position of a lesion in the medical image;
ii) based on the anatomical information, identifying a plurality of anatomical landmarks which have anatomical relations with the lesion;
iii) detecting the plurality of anatomical landmarks in the medical image; and
iv) determining a nearest anatomical landmark of the plurality of anatomical landmarks, wherein determining the nearest anatomical landmark comprises two or more of:
a) determining a plurality of distances between the lesion and individual ones of the plurality of anatomical landmarks, and identifying the nearest anatomical landmark based on a closest distance between the lesion and an individual one of the plurality of anatomical landmarks;
b) determining an anatomical structure that includes the lesion and identifying the nearest anatomical landmark based on the anatomical structure; and
c) identifying the nearest anatomical landmark based on an estimated relevance of the anatomical relations of the plurality of anatomical landmarks to the lesion in a medical interpretation of the lesion; and

a key-image generator for:
j) selecting an image area in the medical image which comprises the lesion and the nearest anatomical landmark of the plurality of anatomical landmarks, and
jj) generating the key-image by cropping or extracting the selected image area from the medical image.

2. The system according to claim 1 , wherein the image analyzer is arranged for:
detecting the anatomical structure in the medical image and identifying the nearest anatomical landmark based on the anatomical structure;
and one or both of
a) determining the plurality of distances between the lesion and the individual ones of the plurality of anatomical landmarks, and identifying the nearest anatomical landmark based on the closest distance between the lesion and the individual one of the plurality of anatomical landmarks; and
c) identifying the nearest anatomical landmark based on the estimated relevance of the anatomical relations of the anatomical landmarks to the lesion in the medical interpretation of the lesion.

3. The system according to claim 2 , wherein the nearest anatomical landmark and an edge of the anatomical structure are included in the image area.

4. The system according to claim 1 , wherein the anatomical information enables the medical interpretation of the lesion, and wherein the image analyzer is arranged for identifying the nearest anatomical landmark based on the estimated relevance of the anatomical relations of the individual landmarks to the lesion in the medical interpretation of the lesion and one or both of:
a) determining a plurality of distances between the lesion and individual ones of the plurality of anatomical landmarks, and identifying the nearest anatomical landmark based on the closest distance between the lesion and the individual one of the plurality of anatomical landmarks; and
b) determining an anatomical structure that includes the lesion and identifying the nearest anatomical landmark based on the anatomical structure.

5. The system according to claim 1 , wherein the image analyzer is arranged for identifying the anatomical landmark based on the distances from individual ones of the plurality of anatomical landmarks to the position of the lesion and one or both of:
b) determining the anatomical structure that includes the lesion and identifying the nearest anatomical landmark based on the anatomical structure; and
c) identifying the nearest anatomical landmark based on the estimated relevance of the anatomical relations of the plurality of anatomical landmarks to the lesion in the medical interpretation of the lesion.

6. The system according to claim 1 , further comprising a user interface subsystem for enabling a user to adjust the image area before the key-image is generated.

7. The system according to claim 1 , wherein the key-image generator is arranged for selecting the image area by establishing a bounding box in the medical image which encompasses the lesion and the nearest anatomical landmark.

8. The system according to claim 7 , wherein the key-image generator is arranged for maintaining an aspect ratio of the bounding box over a number of medical images.

9. The system according to claim 1 , wherein the key-image generator is arranged for adjusting a window level and a window width of the key-image.

10. The system according to claim 1 , wherein the medical database comprises a textbook view of the content of the medical image, and wherein the system is arranged for generating a further key-image from the textbook view which comprises the nearest anatomical landmark and the position of the lesion.

11. The system according to claim 10 , further comprising annotation means for enabling a user to include an annotation in the key-image, and wherein the system is arranged for including the annotation at a corresponding location in the further key-image.

12. The system according to claim 11 , further arranged for registering the textbook view towards the medical image before generating the further key-image.

13. Workstation or imaging apparatus comprising the system of claim 1 .

14. Method for generating a key-image from a medical image, the key-image comprising an image area selected from the medical image, the method comprising:
obtaining the medical image and metadata indicative of a content of the medical image;
accessing anatomical information on a medical database, the anatomical information being indicative of anatomical landmarks associated with the content of the medical image;
detecting a position of a lesion in the medical image;
based on the anatomical information, identifying a plurality of anatomical landmarks which have anatomical relations with the lesion;
detecting the plurality of anatomical landmarks in the medical image;
determining a nearest anatomical landmark of the plurality of anatomical landmarks, wherein determining the nearest anatomical landmark comprises two or more of:
a) determining a plurality of distances between the lesion and individual ones of the plurality of anatomical landmarks, and identifying the nearest anatomical landmark based on a closest distance between the lesion and an individual one of the plurality of anatomical landmarks;
b) determining an anatomical structure that includes the lesion and identifying the nearest anatomical landmark based on the anatomical structure; and
c) identifying the nearest anatomical landmark based on an estimated relevance of the anatomical relations of the plurality of anatomical landmarks to the lesion in a medical interpretation of the lesion; and

selecting an image area in the medical image which comprises the lesion and the nearest anatomical landmark of the plurality of anatomical landmarks; and
generating the key-image by cropping or extracting the selected image area from the medical image.

15. A computer program product stored on a non-transitory computer-readable medium comprising instructions for causing a processor system to perform the method according to claim 14 .

Assignees

• Koninklijke Philips Nv

Inventors

• Serlie Iwo Willem Oscar
• Aleksovski Zarko
• Martherus Rudolph

Classifications

• G06T11/60 Primary

  Creating or editing images; Combining images with text · CPC title

• G06T7/11

  Region-based segmentation · CPC title

• G06T2207/30004

  Biomedical image processing · CPC title

• G06T7/194

  involving foreground-background segmentation · CPC title

• G06T7/0012 Primary

  Biomedical image inspection · CPC title